Happiness (Assignment 1)

Background

* There is no universally accepted definition of happiness, but there have been some ideas about how it
could be measured.
* In this assignment, we investigate the data from the World Happiness Report and produce a report
about this data

Data Source

* World Happiness Data for 2023 : https://worldhappiness.report/data/  : Table 2.1
* Dataset: DataForTable2.1WHR2023.xls

Approach

* We first generate the EDA reports and then the requested questions are answered in the end

References for Conclusion

1. How to Report Skewness and Kurtosis
2. Interpretation of the Kurtosis Coefficient

# Import libraries
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

EDAAnalyzer Class

* The `EDAAnalyzer` class is designed for exploratory data analysis (EDA) on a given DataFrame.

class EDAAnalyzer:
    
    def __init__(self, dataframe):
        self.dataframe = dataframe

    def display_head(self, n=5):
        return self.dataframe.head(n)

    def display_shape(self):
        df_rows, df_cols = self.dataframe.shape
        return "Shape of data: rows: " + str(df_rows) + " cols: " + str(df_cols)

    def display_column_info(self):
        # A replacement for df.info but does the same functionality
        column_names = self.dataframe.columns.tolist()
        df_rows = len(self.dataframe)
        column_info_list = []

        for col in column_names:
            non_null_count = self.dataframe[col].count()
            percent_non_null = (non_null_count / df_rows) * 100
            data_type = self.dataframe[col].dtype

            column_info_list.append({
                'Column Name': col,
                'Non-null Count': non_null_count,
                'Percent Non-null': percent_non_null,
                'Data Type': data_type
            })

        column_info = pd.DataFrame(column_info_list)
        column_info = column_info.sort_values(by='Percent Non-null', ascending=False)
        column_info = column_info.reset_index(drop=True)

        return column_info

    def describe(self):
        return self.dataframe.describe().T

    def plot_histograms(self, ignore_columns=[]):
        columns_to_plot = [col for col in self.dataframe.columns if col not in ignore_columns]
        len_columns = len(columns_to_plot)

        num_rows = (len_columns + 3) // 4
        fig, axs = plt.subplots(nrows=num_rows, ncols=4, figsize=(20, 4 * num_rows))
        axs = axs.flatten()

        for i, col in enumerate(columns_to_plot):
            sns.histplot(self.dataframe[col], ax=axs[i], kde=True)
            
        for j in range(i + 1, len(axs)):
            axs[j].axis('off')
        plt.subplots_adjust(hspace=0.5)
            
        plt.show()

    def get_skew_kurt(self, ignore_columns=[]):
        skew_kurt_data = {'Column Name': [], 'Skewness': [], 'Kurtosis': [], 'Category': []}
        columns_to_analyze = [col for col in self.dataframe.columns if col not in ignore_columns]
        skew_kurt_data = []
        
        for col in columns_to_analyze:
            skewness = abs(self.dataframe[col].skew())
            kurtosis = abs(self.dataframe[col].kurt())
            
            # Ref : https://www.statology.org/how-to-report-skewness-kurtosis/
            category = "Power Data" if skewness > 1 and kurtosis > 3 else \
                       "Normal Data" if abs(skewness) <= 1 and abs(kurtosis) <= 3 else \
                       "Inbetween"
            skew_row_data = {
                'Column Name': col,
                'Skewness': skewness,
                'Kurtosis': kurtosis,
                'Category': category
            }
            
            skew_kurt_data.append(skew_row_data)
        return pd.DataFrame(skew_kurt_data).sort_values(by='Category')



    def get_scatter_plot(self, y_col, add_trend=False):
        numeric_columns = self.dataframe.select_dtypes(include=['number']).columns
        numeric_columns = [col for col in numeric_columns if col not in [y_col]]

        num_plots = len(numeric_columns)
        num_cols = 3
        num_rows = (num_plots + 1) // 2
        fig, axs = plt.subplots(num_rows, num_cols, figsize=(12, 8))
        axs = axs.flatten()
    
        for i, col in enumerate(numeric_columns):
            if(not add_trend):
                sns.scatterplot(data=self.dataframe, x=self.dataframe[col], y=self.dataframe[y_col], ax=axs[i])
            else:    
                sns.regplot(data=self.dataframe, x=self.dataframe[col], y=self.dataframe[y_col], ax=axs[i], scatter=False, color='red')

        for j in range(i + 1, len(axs)):
            axs[j].axis('off')

        plt.subplots_adjust(hspace=0.5)
        plt.show()

    def get_pair_plot(self):
        return sns.pairplot(self.dataframe)

    def get_box_plot(self):
        numeric_columns = self.dataframe.select_dtypes(include=['number']).columns

        num_plots = len(numeric_columns)
        num_cols = 3
        num_rows = (num_plots + 1) // 2
        fig, axs = plt.subplots(num_rows, num_cols, figsize=(20, 14))
        axs = axs.flatten()

        for i, col in enumerate(numeric_columns):
            sns.boxplot(data=self.dataframe, y=self.dataframe[col], ax=axs[i])
            axs[i].set_xlabel(col)

        for j in range(i + 1, len(axs)):
            axs[j].axis('off')

        plt.subplots_adjust(hspace=0.5)
        plt.show()

    def get_heat_map(self):
        non_numeric_columns = self.dataframe.select_dtypes(exclude=['number']).columns
        correlation_matrix = self.dataframe.drop(non_numeric_columns, axis=1).corr().abs()
        sns.heatmap(correlation_matrix, annot=True)
        plt.show()

    def get_correlation_pairs(self):

        # Reference 2
        threshold_high=0.7
        threshold_low=0.3
        
        numeric_columns = self.dataframe.select_dtypes(include=['number']).columns
        correlation_matrix = self.dataframe[numeric_columns].corr().abs()

        correlation_pairs_df = []

        for i,col1 in enumerate(numeric_columns):
            for j,col2 in enumerate(numeric_columns):
                if col1 != col2:
                    if i < j:
                        
                        correlation_coefficient = correlation_matrix.loc[col1, col2]

                        if correlation_coefficient > threshold_high:
                            correlation_category = 'High'
                        elif correlation_coefficient <= threshold_low:
                            correlation_category = 'Low'
                        else:
                            correlation_category = 'Moderate'
                            
                        correlation_pairs_df.append({
                            'Column 1': col1,
                            'Column 2': col2,
                            'Correlation Coefficient': correlation_coefficient,
                            'Correlation Category': correlation_category
                        })

        return pd.DataFrame(correlation_pairs_df)

    
    def get_top_correlated_pairs(self, n=10):
        top_correlated_pairs = self.get_correlation_pairs().sort_values(by='Correlation Coefficient', ascending=False).head(n)
        return top_correlated_pairs

    def get_top_correlated_pairs_by_column(self,column_name,n=10):
        top_correlated_pairs = self.get_correlation_pairs()
        top_correlated_pairs = top_correlated_pairs[(top_correlated_pairs["Column 1"]==column_name) | (top_correlated_pairs["Column 2"]==column_name)]
        return top_correlated_pairs.sort_values(by='Correlation Coefficient', ascending=False).head(n)

    def plot_life_ladder_over_time(self):
        plt.figure(figsize=(12, 6))
        for country in self.dataframe['Country name'].unique():
            country_info = self.dataframe[self.dataframe['Country name'] == country]
            plt.plot(country_info['year'], country_info['Life Ladder'], label=country, linewidth=0.5)

        plt.legend(loc='upper left', bbox_to_anchor=(1, 1), ncol=1)
        plt.show()

    def plot_life_ladder_change_over_time(self):
        plt.figure(figsize=(12, 6))
        change_data = self.dataframe
        change_data['Life Ladder (% Change)'] = change_data.groupby('Country name')['Life Ladder'].pct_change() * 100
        for country in change_data['Country name'].unique():
            country_info = change_data[change_data['Country name'] == country]
            plt.plot(country_info['year'], country_info['Life Ladder (% Change)'], label=country, linewidth=0.2)

        plt.legend(loc='upper left', bbox_to_anchor=(1, 1), ncol=1)
        plt.show()


    def get_missing_rows(self):
        return self.dataframe[self.dataframe.isnull().any(axis=1)]

    def get_most_least_happy_countries(self, years):
        result = []

        for year in years:
            year_data = self.dataframe[self.dataframe['year'] == year]

            most_happy = year_data.loc[year_data['Life Ladder'].idxmax()]
            least_happy = year_data.loc[year_data['Life Ladder'].idxmin()]

            result.append({
                'Year': year,
                'Most Happy Country': most_happy['Country name'],
                'Most Happy Value': most_happy['Life Ladder'],
                'Least Happy Country': least_happy['Country name'],
                'Least Happy Value': least_happy['Life Ladder']
            })

        most_happy_all_time = self.dataframe.loc[self.dataframe['Life Ladder'].idxmax()]
        least_happy_all_time = self.dataframe.loc[self.dataframe['Life Ladder'].idxmin()]

        result.append({
            'Year': 'All Time',
            'Most Happy Country': most_happy_all_time['Country name'],
            'Most Happy Value': most_happy_all_time['Life Ladder'],
            'Least Happy Country': least_happy_all_time['Country name'],
            'Least Happy Value': least_happy_all_time['Life Ladder']
        })

        result_df = pd.DataFrame(result)
        return result_df


    def find_largest_change_in_happiness(self,current_year=2022):

        result = []
        for country in self.dataframe['Country name'].unique():
            country_info = self.dataframe[self.dataframe['Country name'] == country]
            min_row = country_info.loc[country_info['year'].idxmin()]
            max_row = country_info.loc[country_info['year'].idxmax()]

            if(max_row["year"]==current_year):
                result.append({
                    'Country': country,
                    'Start Year': min_row["year"],
                    'Start Happiness': min_row["Life Ladder"],
                    'End Year': max_row["year"],
                    'End Happiness': max_row["Life Ladder"],
                    '% Change': 100*((max_row["Life Ladder"]-min_row["Life Ladder"])/(min_row["Life Ladder"]))
                    })

        return pd.DataFrame(result)
        

# Load data
data_loc = "DataForTable2.1WHR2023.xls"
df = pd.read_excel(data_loc)

# Create an instance of EDAAnalyzer
eda_analyzer = EDAAnalyzer(df)

eda_analyzer.display_head(10)

	Country name	year	Life Ladder	Log GDP per capita	Social support	Healthy life expectancy at birth	Freedom to make life choices	Generosity	Perceptions of corruption	Positive affect	Negative affect
0	Afghanistan	2008	3.723590	7.350416	0.450662	50.500000	0.718114	0.167652	0.881686	0.414297	0.258195
1	Afghanistan	2009	4.401778	7.508646	0.552308	50.799999	0.678896	0.190809	0.850035	0.481421	0.237092
2	Afghanistan	2010	4.758381	7.613900	0.539075	51.099998	0.600127	0.121316	0.706766	0.516907	0.275324
3	Afghanistan	2011	3.831719	7.581259	0.521104	51.400002	0.495901	0.163571	0.731109	0.479835	0.267175
4	Afghanistan	2012	3.782938	7.660506	0.520637	51.700001	0.530935	0.237588	0.775620	0.613513	0.267919
5	Afghanistan	2013	3.572100	7.680333	0.483552	52.000000	0.577955	0.062666	0.823204	0.547417	0.273328
6	Afghanistan	2014	3.130896	7.670638	0.525568	52.299999	0.508514	0.105755	0.871242	0.491641	0.374861
7	Afghanistan	2015	3.982855	7.653833	0.528597	52.599998	0.388928	0.081652	0.880638	0.491410	0.339276
8	Afghanistan	2016	4.220169	7.650370	0.559072	52.924999	0.522566	0.043916	0.793246	0.501409	0.348332
9	Afghanistan	2017	2.661718	7.647830	0.490880	53.250000	0.427011	-0.119410	0.954393	0.435270	0.371326

Info on dataset

# Data shape
eda_analyzer.display_shape()

'Shape of data: rows: 2199 cols: 11'

# Display null info and type for columns
# below function replaces df.info

eda_analyzer.display_column_info()

	Column Name	Non-null Count	Percent Non-null	Data Type
0	Country name	2199	100.000000	object
1	year	2199	100.000000	int64
2	Life Ladder	2199	100.000000	float64
3	Social support	2186	99.408822	float64
4	Negative affect	2183	99.272397	float64
5	Log GDP per capita	2179	99.090496	float64
6	Positive affect	2175	98.908595	float64
7	Freedom to make life choices	2166	98.499318	float64
8	Healthy life expectancy at birth	2145	97.544338	float64
9	Generosity	2126	96.680309	float64
10	Perceptions of corruption	2083	94.724875	float64

eda_analyzer.describe()

	count	mean	std	min	25%	50%	75%	max
year	2199.0	2014.161437	4.718736	2005.000000	2010.000000	2014.000000	2018.000000	2022.000000
Life Ladder	2199.0	5.479226	1.125529	1.281271	4.646750	5.432437	6.309460	8.018934
Log GDP per capita	2179.0	9.389766	1.153387	5.526723	8.499764	9.498955	10.373216	11.663788
Social support	2186.0	0.810679	0.120952	0.228217	0.746609	0.835535	0.904792	0.987343
Healthy life expectancy at birth	2145.0	63.294583	6.901104	6.720000	59.119999	65.050003	68.500000	74.474998
Freedom to make life choices	2166.0	0.747858	0.140150	0.257534	0.656528	0.769821	0.859382	0.985178
Generosity	2126.0	0.000096	0.161083	-0.337527	-0.112116	-0.022671	0.092070	0.702708
Perceptions of corruption	2083.0	0.745195	0.185837	0.035198	0.688139	0.799654	0.868827	0.983276
Positive affect	2175.0	0.652143	0.105922	0.178886	0.571684	0.663063	0.737936	0.883586
Negative affect	2183.0	0.271501	0.086875	0.082737	0.207660	0.260671	0.322894	0.704590

Univariate Analysis

eda_analyzer.plot_histograms(ignore_columns=["Country name"])

eda_analyzer.get_box_plot()

/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/seaborn/categorical.py:640: FutureWarning: SeriesGroupBy.grouper is deprecated and will be removed in a future version of pandas.
  positions = grouped.grouper.result_index.to_numpy(dtype=float)
/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/seaborn/categorical.py:640: FutureWarning: SeriesGroupBy.grouper is deprecated and will be removed in a future version of pandas.
  positions = grouped.grouper.result_index.to_numpy(dtype=float)
/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/seaborn/categorical.py:640: FutureWarning: SeriesGroupBy.grouper is deprecated and will be removed in a future version of pandas.
  positions = grouped.grouper.result_index.to_numpy(dtype=float)
/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/seaborn/categorical.py:640: FutureWarning: SeriesGroupBy.grouper is deprecated and will be removed in a future version of pandas.
  positions = grouped.grouper.result_index.to_numpy(dtype=float)
/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/seaborn/categorical.py:640: FutureWarning: SeriesGroupBy.grouper is deprecated and will be removed in a future version of pandas.
  positions = grouped.grouper.result_index.to_numpy(dtype=float)
/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/seaborn/categorical.py:640: FutureWarning: SeriesGroupBy.grouper is deprecated and will be removed in a future version of pandas.
  positions = grouped.grouper.result_index.to_numpy(dtype=float)
/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/seaborn/categorical.py:640: FutureWarning: SeriesGroupBy.grouper is deprecated and will be removed in a future version of pandas.
  positions = grouped.grouper.result_index.to_numpy(dtype=float)
/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/seaborn/categorical.py:640: FutureWarning: SeriesGroupBy.grouper is deprecated and will be removed in a future version of pandas.
  positions = grouped.grouper.result_index.to_numpy(dtype=float)
/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/seaborn/categorical.py:640: FutureWarning: SeriesGroupBy.grouper is deprecated and will be removed in a future version of pandas.
  positions = grouped.grouper.result_index.to_numpy(dtype=float)
/Library/Frameworks/Python.framework/Versions/3.11/lib/python3.11/site-packages/seaborn/categorical.py:640: FutureWarning: SeriesGroupBy.grouper is deprecated and will be removed in a future version of pandas.
  positions = grouped.grouper.result_index.to_numpy(dtype=float)

## Bivariate Analysis

eda_analyzer.get_scatter_plot("Life Ladder")

eda_analyzer.get_scatter_plot("Life Ladder",add_trend=True)

eda_analyzer.get_pair_plot()

<seaborn.axisgrid.PairGrid at 0x14f822310>

eda_analyzer.get_heat_map()

eda_analyzer.get_correlation_pairs().head(5)

	Column 1	Column 2	Correlation Coefficient	Correlation Category
0	year	Life Ladder	0.045943	Low
1	year	Log GDP per capita	0.077772	Low
2	year	Social support	0.029750	Low
3	year	Healthy life expectancy at birth	0.163500	Low
4	year	Freedom to make life choices	0.234105	Low

eda_analyzer.get_top_correlated_pairs()

	Column 1	Column 2	Correlation Coefficient	Correlation Category
18	Log GDP per capita	Healthy life expectancy at birth	0.818126	High
9	Life Ladder	Log GDP per capita	0.784871	High
10	Life Ladder	Social support	0.721663	High
11	Life Ladder	Healthy life expectancy at birth	0.713493	High
17	Log GDP per capita	Social support	0.683619	Moderate
24	Social support	Healthy life expectancy at birth	0.597682	Moderate
37	Freedom to make life choices	Positive affect	0.578752	Moderate
12	Life Ladder	Freedom to make life choices	0.534532	Moderate
15	Life Ladder	Positive affect	0.518207	Moderate
36	Freedom to make life choices	Perceptions of corruption	0.476537	Moderate

eda_analyzer.get_top_correlated_pairs_by_column(column_name="Life Ladder")

	Column 1	Column 2	Correlation Coefficient	Correlation Category
9	Life Ladder	Log GDP per capita	0.784871	High
10	Life Ladder	Social support	0.721663	High
11	Life Ladder	Healthy life expectancy at birth	0.713493	High
12	Life Ladder	Freedom to make life choices	0.534532	Moderate
15	Life Ladder	Positive affect	0.518207	Moderate
14	Life Ladder	Perceptions of corruption	0.431569	Moderate
16	Life Ladder	Negative affect	0.339992	Moderate
13	Life Ladder	Generosity	0.181658	Low
0	year	Life Ladder	0.045943	Low

Time Series Analysis

eda_analyzer.plot_life_ladder_over_time()

eda_analyzer.plot_life_ladder_change_over_time()

Questions Answered

1. How large is the dataset

eda_analyzer.display_shape()

'Shape of data: rows: 2199 cols: 12'

2. What are the features , how are they represented?

eda_analyzer.display_column_info()

	Column Name	Non-null Count	Percent Non-null	Data Type
0	Country name	2199	100.000000	object
1	year	2199	100.000000	int64
2	Life Ladder	2199	100.000000	float64
3	Social support	2186	99.408822	float64
4	Negative affect	2183	99.272397	float64
5	Log GDP per capita	2179	99.090496	float64
6	Positive affect	2175	98.908595	float64
7	Freedom to make life choices	2166	98.499318	float64
8	Healthy life expectancy at birth	2145	97.544338	float64
9	Generosity	2126	96.680309	float64
10	Perceptions of corruption	2083	94.724875	float64
11	Life Ladder (% Change)	2034	92.496589	float64

3. Are there any interesting distributions in the (numeric) features?

eda_analyzer.plot_histograms()

- We see a drastic drop in participating countries in 2020. Probably an  impact of COVID. But that number hasnt gone up since then showing a some countries might not be willing to participate in the analytics possibly with disagreements. Needs further readings

eda_analyzer.get_skew_kurt(ignore_columns=["Country name"])

	Column Name	Skewness	Kurtosis	Category
3	Social support	1.118834	1.173909	Inbetween
4	Healthy life expectancy at birth	1.145454	2.989660	Inbetween
7	Perceptions of corruption	1.490418	1.814231	Inbetween
10	Life Ladder (% Change)	0.175574	3.960590	Inbetween
0	year	0.076683	1.071694	Normal Data
1	Life Ladder	0.017819	0.591848	Normal Data
2	Log GDP per capita	0.335130	0.773036	Normal Data
5	Freedom to make life choices	0.670538	0.034785	Normal Data
6	Generosity	0.777087	0.831590	Normal Data
8	Positive affect	0.436205	0.207818	Normal Data
9	Negative affect	0.731083	0.757400	Normal Data

- Social support,Healthy life expectancy at birth,Perceptions of corruption fall neither between normal or a poer distribution, the rest of the variables follow more of a normal distribution

df.groupby('year')['Country name'].nunique().reset_index()

	year	Country name
0	2005	27
1	2006	89
2	2007	102
3	2008	110
4	2009	114
5	2010	124
6	2011	146
7	2012	141
8	2013	136
9	2014	144
10	2015	142
11	2016	141
12	2017	147
13	2018	141
14	2019	143
15	2020	116
16	2021	122
17	2022	114

- Life ladder follows a normal distribution, which is an expected behaviour of a generated value of rating type

- The number of particilating countries from the lower GDP/capita section is very low, resulting in possible not considering parameters of happiness in such countries

- The data trends across different parameters suggest the same that the countries under participation are mostly from the developing or developed spectrum. The data doesnt show uniform distribution across parameters.

- The most skewed data of all is social security. Either coutries offering social sequrity are low or refusal in participation. Needs further reading

4. Are there any interesting correlations in the data?

eda_analyzer.get_heat_map()

eda_analyzer.get_top_correlated_pairs()

	Column 1	Column 2	Correlation Coefficient	Correlation Category
20	Log GDP per capita	Healthy life expectancy at birth	0.818126	High
10	Life Ladder	Log GDP per capita	0.784871	High
11	Life Ladder	Social support	0.721663	High
12	Life Ladder	Healthy life expectancy at birth	0.713493	High
19	Log GDP per capita	Social support	0.683619	Moderate
27	Social support	Healthy life expectancy at birth	0.597682	Moderate
42	Freedom to make life choices	Positive affect	0.578752	Moderate
13	Life Ladder	Freedom to make life choices	0.534532	Moderate
16	Life Ladder	Positive affect	0.518207	Moderate
41	Freedom to make life choices	Perceptions of corruption	0.476537	Moderate

- In general : The most correlated values are Log GDP per capita	and Healthy life expectancy at birth. The rest shows how happiness comes if you are given more money and social support. 

eda_analyzer.get_top_correlated_pairs_by_column(column_name="Life Ladder")

	Column 1	Column 2	Correlation Coefficient	Correlation Category
10	Life Ladder	Log GDP per capita	0.784871	High
11	Life Ladder	Social support	0.721663	High
12	Life Ladder	Healthy life expectancy at birth	0.713493	High
13	Life Ladder	Freedom to make life choices	0.534532	Moderate
16	Life Ladder	Positive affect	0.518207	Moderate
15	Life Ladder	Perceptions of corruption	0.431569	Moderate
17	Life Ladder	Negative affect	0.339992	Moderate
14	Life Ladder	Generosity	0.181658	Low
18	Life Ladder	Life Ladder (% Change)	0.170925	Low
0	year	Life Ladder	0.045943	Low

- Speaking in terms of happiness, as years change things dont change/reflect on happiness as its the least correlated. 
- all parameters are typical in correlation with the standared of living. which I belive is the main parameter the authors consider for happiness.

5. Can you identify any limitations, missing values or distortions of the data?

eda_analyzer.display_column_info()

	Column Name	Non-null Count	Percent Non-null	Data Type
0	Country name	2199	100.000000	object
1	year	2199	100.000000	int64
2	Life Ladder	2199	100.000000	float64
3	Social support	2186	99.408822	float64
4	Negative affect	2183	99.272397	float64
5	Log GDP per capita	2179	99.090496	float64
6	Positive affect	2175	98.908595	float64
7	Freedom to make life choices	2166	98.499318	float64
8	Healthy life expectancy at birth	2145	97.544338	float64
9	Generosity	2126	96.680309	float64
10	Perceptions of corruption	2083	94.724875	float64
11	Life Ladder (% Change)	2034	92.496589	float64

eda_analyzer.get_missing_rows()

	Country name	year	Life Ladder	Log GDP per capita	Social support	Healthy life expectancy at birth	Freedom to make life choices	Generosity	Perceptions of corruption	Positive affect	Negative affect	Life Ladder (% Change)
0	Afghanistan	2008	3.723590	7.350416	0.450662	50.500000	0.718114	0.167652	0.881686	0.414297	0.258195	NaN
13	Afghanistan	2022	1.281271	NaN	0.228217	54.875000	0.368377	NaN	0.733198	0.205868	0.575512	-47.403408
14	Albania	2007	4.634252	9.121704	0.821372	66.760002	0.528605	-0.010429	0.874700	0.488819	0.246335	NaN
29	Algeria	2010	5.463567	9.306355	NaN	65.500000	0.592696	-0.209753	0.618038	NaN	NaN	NaN
32	Algeria	2014	6.354898	9.355415	0.818189	65.900002	NaN	NaN	NaN	0.558359	0.176866	13.387278
...	...	...	...	...	...	...	...	...	...	...	...	...
2155	Yemen	2007	4.477133	8.211859	0.824969	58.720001	0.672685	0.009045	NaN	0.523652	0.378784	NaN
2163	Yemen	2016	3.825631	7.552322	0.775407	58.174999	0.532964	-0.140766	NaN	0.401007	0.227925	28.261790
2164	Yemen	2017	3.253560	7.243477	0.789555	57.950001	0.595191	-0.124107	NaN	0.368106	0.295064	-14.953634
2167	Zambia	2006	4.824455	7.834035	0.797665	46.759998	0.720972	-0.008994	0.785281	0.663519	0.226278	NaN
2182	Zimbabwe	2006	3.826268	7.459545	0.821656	40.400002	0.431110	-0.059926	0.904757	0.668662	0.297147	NaN

352 rows × 12 columns

eda_analyzer.get_missing_rows().drop(columns=['Country name']).mean()

year                                2011.110795
Life Ladder                            5.476915
Log GDP per capita                     9.531857
Social support                         0.824012
Healthy life expectancy at birth      62.653574
Freedom to make life choices           0.749911
Generosity                             0.003443
Perceptions of corruption              0.729850
Positive affect                        0.643346
Negative affect                        0.256682
Life Ladder (% Change)                -0.256926
dtype: float64

eda_analyzer.describe()['mean']

year                                2014.161437
Life Ladder                            5.479226
Log GDP per capita                     9.389766
Social support                         0.810679
Healthy life expectancy at birth      63.294583
Freedom to make life choices           0.747858
Generosity                             0.000096
Perceptions of corruption              0.745195
Positive affect                        0.652143
Negative affect                        0.271501
Life Ladder (% Change)                 0.418340
Name: mean, dtype: float64

- missing values or distortions - 5% data on the perception of corruption is missing and generosity. they also dont have much correlation with happiness which might be a reason why the authors might have been fine with it
  - Reason for missing values should be read upon.

  - Another distortion is the drop in countries participating recently
  

6. What would you like to see in this dataset?

* Employment
* mental health index
* income, gender and race inequality
* leisure time spent
* environmental factors
* Social engagement
* access to healthcare
* their own perception of happiness rating
* internet access
* Safety
* public service
* social media usage
* PPP
* access to higher education
* class mobility
* food sequrity
* housing parameters
* age, average savings, income, spends

7. What is your evaluation of the quality of the data?

In summary, while the dataset provides valuable insights into happiness determinants, addressing missing values, investigating data trends, and incorporating additional factors would further improve its quality and robustness.

The analysis and conclusions generated above answer this question

Data quality evaluvation :

• the dataset describes happiness of different countries and some parameters associated with it over time
• The data has missing values for some columns at 5% max. the missing percent are displayed above
• Can comment much on data outliers are the columns are metrics
• Ther is high covariance inbelow
  • Log GDP per capita Healthy life
  • Life Ladder Log GDP per capita
  • Life Ladder Social support
  • Life Ladder Healthy life expectancy at birth
• Recommendations :
  • consider other parameters that are more related to class, environment and living conditions
  • Some parameters are corelated which can skew the happiness rating. need to declutter them
  • There is no need of any major cleaning as this data is published clean

8. What is the happiest country in 2023? In 2013? Of all time? What is the least happy country in 2023, 2013 and of all time?

eda_analyzer.get_most_least_happy_countries([2013,2022])

	Year	Most Happy Country	Most Happy Value	Least Happy Country	Least Happy Value
0	2013	Canada	7.593794	Syria	2.687553
1	2022	Finland	7.728998	Afghanistan	1.281271
2	All Time	Denmark	8.018934	Afghanistan	1.281271

Happiest country in 2022: Finland (7.728998	).
Happiest country in 2013: Canada (7.593794).
Happiest country of all time: Denmark (8.018934).
Least happy country in 2022: Afghanistan (1.281271).
Least happy country in 2013: Syria (2.687553).
Least happy country of all time: Afghanistan (1.2812).

9. Of the countries in this dataset, which country had the largest increase in happiness from its start of participation to 2023? Which had the largest decrease in happiness?

change_data = eda_analyzer.find_largest_change_in_happiness().sort_values(by="% Change")
change_data

	Country	Start Year	Start Happiness	End Year	End Happiness	% Change
0	Afghanistan	2008	3.723590	2022	1.281271	-65.590435
61	Lebanon	2005	5.491245	2022	2.352428	-57.160396
55	Jordan	2005	6.294660	2022	4.355607	-30.804738
93	Sierra Leone	2006	3.628185	2022	2.560430	-29.429465
10	Botswana	2006	4.739367	2022	3.435275	-27.516163
...	...	...	...	...	...	...
103	Togo	2006	3.202429	2022	4.238982	32.367706
12	Bulgaria	2007	3.843798	2022	5.378349	39.922778
80	Nicaragua	2006	4.460158	2022	6.392258	43.319075
37	Georgia	2006	3.675108	2022	5.292755	44.016298
20	Congo (Brazzaville)	2008	3.819792	2022	5.804919	51.969488

114 rows × 6 columns

largest_increase = change_data.loc[change_data['% Change'].idxmax()]

print(str(largest_increase['Country'])+" got the largest increase in happiness from "+str(largest_increase['Start Happiness'])+" to "+str(largest_increase['End Happiness'])+ " from "+ str(largest_increase['Start Year'])+ " to "+str(largest_increase['End Year']))

Congo (Brazzaville) got the largest increase in happiness from 3.8197922706604004 to 5.8049187660217285 from 2008 to 2022

largest_increase = change_data.loc[change_data['% Change'].idxmin()]

print(str(largest_increase['Country'])+" got the largest decrease in happiness from "+str(largest_increase['Start Happiness'])+" to "+str(largest_increase['End Happiness'])+ " from "+ str(largest_increase['Start Year'])+ " to "+str(largest_increase['End Year']))

Afghanistan got the largest decrease in happiness from 3.7235898971557617 to 1.2812711000442505 from 2008 to 2022